Designing autonomous agents has been one of the major concerns in several fields of research. In software engineering, intelligent computer programs have been developed that can be viewed as autonomous software agents. These programs automatically accomplish various software tasks with little intervention from the user. In robotics, current research has concentrated on developing autonomous mobile robots that can function successfully with little or no human monitoring. In computer graphics, by modeling each animated character as an autonomous agent, complex animations can be produced with minimal intervention from the animator, as is demonstrated in this thesis.
Any autonomous agent will encounter the problem of action selection. The task of action selection (also known as behavior arbitration or behavioral choice) is to determine, from a set of available actions, the most appropriate one based on the agent's internal and external conditions. Designing effective action selection mechanisms is a major endeavor in the design of autonomous agents. To this end, two questions need to be answered first: ``what do we mean by an action?'' and ``what do we mean by the most appropriate action?''.
|Xiaoyuan Tu||January 1996|